Application Refactoring · IT Operations Automation
NIO wanted to raise software-engineering speed for smart-cockpit work without giving up code quality or security. The team needed help with repetitive coding, interface lookup, testing, and bug prevention, but it also required enterprise controls around source-code privacy and tailored coding standards.
Reported impact
NIO embedded AI directly into engineering rather than running isolated code experiments. The team now uses AI-generated code and tests at meaningful scale, supports faster smart-cockpit delivery, and keeps source-code assistance inside a governed enterprise environment built around quality and security requirements.
NIO introduced Alibaba Cloud Qoder CN into its smart-cockpit engineering workflow as a governed AI software-factory layer. Developers use the assistant for code completion, natural-language-to-code generation, interface interpretation, comment generation, and knowledge-backed code templates so they can spend less time on repetitive tasks and uncommon API lookups. The platform also includes a unit-test agent that can generate tests against changed files, compile and run them, and help repair failures, which turns AI into part of the validation path instead of only a code suggester. NIO also uses Qoder CN's knowledge-management and extension-management capabilities to codify naming rules, API-document generation, and historical bug-prevention prompts under enterprise governance. The deployment remains inside a dedicated enterprise environment so source code stays under internal control while the team still gets broad AI assistance across development and maintenance.
Developers invoke Qoder CN inside their daily engineering workflow, where project context, prompts, and enterprise knowledge are used to generate code, comments, tests, and risk warnings. Outputs flow back into the development process and are refined with NIO's extension and knowledge rules before code is committed.
In the smart-cockpit team, nearly 400 people use Qoder CN and AI-generated code can exceed 70% of incremental code in the Tiantan project. (original: 已有近 400 名成员在工作中使用 Qoder CN... 业务增量代码 AI 生成占比最高可达 70% 以上。)
NIO said Qoder CN helps developers generate code, explain interfaces, add comments, and create unit tests while meeting enterprise security requirements. (original: 阿里云 Qoder CN 企业专属版不仅可以通过代码智能生成、研发智能问答、多文件代码修改、自主执行等能力... 还可以满足企业安全合规要求。)
The team uses knowledge management and custom extensions for standards checks, variable naming, API document generation, and RAG-style bug prevention. (original: 利用自定义扩展指令... 代码规范检查、代码变量命名以及 API 开发文档自动生成... 将团队编程规范、过往经常出现的 Bug 等数据以 Prompt、RAG 方式托管至 Qoder CN 企业后台。)
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Alibaba Cloud Sample Sep 20, 2026 Fintopia (瓴岳科技) Fintopia had already embedded AI skills across acquisition, risk review, settlement, operations, and internal support, but hundreds of skills were becoming fragmented assets. Teams kept rebuilding similar capabilities,… connected 50+ models AgentLoop · LoongSuite Collector · MCP Plaza Alibaba Cloud Sample Jul 28, 2026 Zhuiyin Technology (追因科技) Zhuiyin Technology had already built an agent for cross-border ecommerce analysis, but once the agent was exposed to customers, the team discovered three hard problems: results were hard to judge objectively, reasoning… inference quality improved 20% AgentLoop · Trekin Agent Alibaba Cloud Sample Jun 9, 2026 AsiaInfo Technologies (亚信科技) AsiaInfo had to raise software-engineering productivity while maintaining large, long-lived telecom BSS and OSS systems whose logic spanned multiple generations from 2G to 5G. The codebase and architecture were highly… average end-to-end efficiency from requirements to testing improved 8.6% Qoder · Qoder CN 